Métodos de classificação por teoria da decisão para mensuração de dados do ENEM.

Abstract

In this work, the decision tree algorithm was used to analyze the relation of the essay grade with socioeconomic variables of candidates who took the ENEM 2019 and declared residence in the city of Cuité - PB. Our objective was to generate a decision tree using the ranking algorithm, based on which we could find factors that are important for the candidates' performance in the essay. We specifically achieved this goal by building a decision model, whose steps included data collection and preparation, data pre-processing (data cleaning, conversion), model building (algorithm training) and algorithm optimization for model simulation. In this scenario, only three variables were considered significant in the decision tree plot, which were marital status, age and home computer with a 75% accuracy. Given the above, it is believed that other models should be validated in future work seeking to improve this accuracy by inserting new variables into analysis.